[ovs-dev] [PATCH 1/1] vswitchd: Allow setting MAC on DPDK interfaces

Ilya Maximets i.maximets at ovn.org
Fri Jan 3 14:56:59 UTC 2020

On 23.12.2019 22:34, Eveline Raine wrote:
> When setting mac address for an interface using ovs-vsctl command:
>     ovs-vsctl set interface <iface> mac=XX:XX:XX:XX:XX:XX
> iface_set_mac() is responsible to delegate a request to set MAC to a
> netdev-specific set_etheraddr().
> At the moment iface_set_mac() skips all interfaces except those with
> type = "internal", making it impossible to change MAC on any DPDK port.
> Since DPDK ports are owned by the OVS process, OVSDB can be considered
> the source of truth for them. In particular, the source of truth for
> their MAC addresses - so, OVS can take responsibility for setting them.
> Therefore this check is extended to "dpdk" type.
> Acked-by: Roni Bar Yanai <roniba at mellanox.com>
> Tested-by: Adrian Chiris <adrianc at mellanox.com>
> Signed-off-by: Eveline Raine <eveliner at mellanox.com>
> ---
>  vswitchd/bridge.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> diff --git a/vswitchd/bridge.c b/vswitchd/bridge.c
> index 5de0a264a..355364afd 100644
> --- a/vswitchd/bridge.c
> +++ b/vswitchd/bridge.c
> @@ -4696,7 +4696,7 @@ iface_set_mac(const struct bridge *br, const struct port *port, struct iface *if
>      struct eth_addr ea, *mac = NULL;
>      struct iface *hw_addr_iface;
> -    if (strcmp(iface->type, "internal")) {
> +    if (strcmp(iface->type, "internal") && strcmp(iface->type, "dpdk")) {

Hmm... Interesting.  I didn't look at this code closely before and it
actually allows setting mac only for internal ports.

Allowing only dpdk port types here doesn't seem right.  Why not allowing
all the port types?

The issue I can see in this code is that OVS inherits MACs for internal
ports from the physical ones, i.e. if we're changing the MAC of a physical
port we need to re-check all the internal ports and update their addresses.

This might happen almost automatically on linux since we probably will
receive if-notifier event about MAC changing and OVS will trigger re-applying
of the bridge configuration.  However, this will not happen in DPDK case.
Recently introduced manual if-notifier could help here, I guess.
But all of this needs to be tested.

Ben, do you see any other drawbacks that we should handle if we'll allow
changing MAC addresses for non-internal ports?  Or, maybe some issues with
my logic?

Best regards, Ilya Maximets.

More information about the dev mailing list